climetlab-s2s-ai-challenge


Nameclimetlab-s2s-ai-challenge JSON
Version 0.9.0 PyPI version JSON
download
home_pagehttps://github.com/ecmwf-lab/climetlab-s2s-ai-challenge
SummaryClimetlab external dataset plugin for the S2S AI competition organised by ECMWF
upload_time2023-09-13 08:11:43
maintainer
docs_urlNone
authorEuropean Centre for Medium-Range Weather Forecasts (ECMWF)
requires_python
licenseApache License Version 2.0
keywords meteorology
VCS
bugtrack_url
requirements No requirements were recorded.
Travis-CI No Travis.
coveralls test coverage No coveralls.
            [![Check and publish Python Package](https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/actions/workflows/check-and-publish.yml/badge.svg)](https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/actions/workflows/check-and-publish.yml) 
[![PyPI version fury.io](https://badge.fury.io/py/climetlab-s2s-ai-challenge.svg)](https://pypi.python.org/pypi/climetlab-s2s-ai-challenge/)
[![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/ecmwf-lab/climetlab-s2s-ai-challenge/main?urlpath=lab)

# S2S AI challenge Datasets

This is a `climetlab` plugin for Sub-seasonal to Seasonal (S2S) Artificial Intelligence Challenge: https://s2s-ai-challenge.github.io/.

In this README is a description of how to get the data for the S2S AI challenge. Here is a more general [description of the S2S data](https://confluence.ecmwf.int/display/S2S/Description). The data used for the S2S AI challenge is a subset of the S2S data library. More detail can be found at https://confluence.ecmwf.int/display/S2S  and https://confluence.ecmwf.int/display/S2S/Parameters.

There are several ways to use the datasets. Either by direct download (`wget`, `curl`, `browser`) for [`GRIB`](https://en.wikipedia.org/wiki/GRIB) and [NetCDF](https://en.wikipedia.org/wiki/NetCDF) formats; or using the `climetlab` python package with this addon, for `GRIB` and `NetCDF` and `zarr` formats. [`zarr`](https://zarr.readthedocs.io/en/stable/) is a cloud-friendly experimental data format and supports dowloading only the part of the data that is required. It has been designed to work better than classical format on a cloud environment (experimental).

# Installation

`pip install -U climetlab climetlab_s2s_ai_challenge`


# API

```python
import climetlab
```

Use `climetlab.load_dataset('s2s-ai-challenge-{datasetname}')` with the following keywords:

- `datasetname`: name of the dataset, see [dataset description](#datasets-description)
- `parameter`: [variable](#parameter), see [hindcast input for the different models](#hindcast-input)
- `origin`: name of the model [`ecmwf`, `eccc`, `ncep`] or modelling center [`ecmf`, `cwao`, `kwbc`]. Provide `origin` only for `training/test-input`/`hindcast/forecast-input`.
- `date`: `YYYYMMDD` is the date of the 2020 forecast for `test-input`/`forecast-input`. The same dates are required for the on-the-fly `training-input`/`hindcast-input` but returns the multi-year hindcast for the given `MMDD` date. Please provide `int`, `str`, `np.datetime` or list of the former in format `YYYYMMDDD`. Providing no `date` keyword downloads all dates.
- `format`: data format, choose from [`netcdf` (always available), `grb` (only for `input-*`), `zarr` (experimental)]

## Coordinates

Overview of the time-related coordinates.

| `name` | CF convention `standard_name` | description | comment |
| --- | --- | --- | --- |
| `forecast_time` | forecast_reference_time | The forecast reference time in NWP is the "data time", the time of the analysis from which the forecast was made. It is not the time for which the forecast is valid. | |
| `lead_time` | forecast_period | Forecast period is the time interval between the forecast reference time and the validity time. |  |
| `valid_time` | time | time for which the forecast is valid | `forecast_time` + `lead_time` |

All datasets are on a global 1.5 degree grid.


## Parameter

`parameter` describes the variable to download. The most important two variables for the `s2s-ai-challenge` are the two target variables `t2m` and `tp`:

| parameter | long_name | standard_name | unit | description & aggregation type | week 3-4 aggregation | week 5-6 aggregation | link to source |
| -------- | --------- | --- | --- | --- | --- | --- | --- |
| `t2m`    | 2m temperature | air_temperature | K | Temperature at 2m height averaged for the date given | average [day 14, day 27] | average [day 28, day 41] | [model](https://confluence.ecmwf.int/display/S2S/S2S+Surface+Air+Temperature), [observations](http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NCEP/.CPC/.temperature/.daily/)|
| `tp`     | total precipitation | precipitation_amount | kg m-2 | Total precipitation accumulated from `forecast_time` until including `valid_time`, e.g. `lead_time = 1 days` accumulates precipitation_flux `pr` from 6-hourly steps 0,6,12,18 at date `forecast_time` | day 28 minus day 14 | day 42 minus day 28 | [model](https://confluence.ecmwf.int/display/S2S/S2S+Total+Precipitation) | 
| `pr`     | precipitation flux | precipitation_flux | kg m-2 | Precipitation accumulated for the date given | use `tp` | use `tp` | [observations](http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NCEP/.CPC/.UNIFIED_PRCP/.GAUGE_BASED/.GLOBAL/.v1p0/.extREALTIME/.rain) |

Given the different nature of the parameters, with `forecast_time` Jan 2nd 2020, `tp` `lead_time=1 days` and `t2m` `lead_time=0 days` describe both the weather of Jan 2nd 2020, as `tp` is aggregated `pr` from Jan 2nd 00:00 to Jan 3rd 00:00 and `t2m` as the average of Jan 2nd. Furthermore `tp` is aggregated since `forecast_time`, i.e. `tp` `lead_time=5 days` is `pr` aggregated from Jan 2nd 00:00 to Jan 7th 00:00.

For the remaining variable descriptions, see [ECWMF S2S description](https://confluence.ecmwf.int/display/S2S/Parameters).


## Datasets description

There are four datasets provided for this challenge. As we are aiming at bringing together the two communities of Machine Learning and Weather Prediction, they have been aliases to use both two points of views:

| ML                          | NWP                          |                                                        |
| --------------------------- | ---------------------------- | ------------------------------------------------------ |
| `training-input`            | `hindcast-input`             | Training  dataset (input for training the ML models)   |
| `training-output-reference` | `hindcast-like-observations` | Training dataset (output for training the ML models)   |
| `training-output-benchmark` | `hindcast-benchmark`         | Benchmark output (on the training dataset)             |
| `test-input`                | `forecast-input`             | Test dataset (DO NOT use for training)                 |
| `test-output-reference`     | `forecast-like-observations` | Test dataset (DO NOT use)                              |
| `test-output-benchmark`     | `forecast-benchmark`         | Benchmark output (on the test dataset)                 |
| `observations`              | `observations`               | Observations with `time` dimension                     |


**Overfitting** is always an potential issue when using ML algorithms. To address this, the data is usually split into three datasets : 
[training](https://en.wikipedia.org/wiki/Training,_validation,_and_test_sets#Training_dataset), 
[validation](https://en.wikipedia.org/wiki/Training,_validation,_and_test_sets#Validation_dataset) 
and [test](https://en.wikipedia.org/wiki/Training,_validation,_and_test_sets#Test_dataset). 
This terminology has lead to [some confusion in the past](https://en.wikipedia.org/wiki/Training,_validation,_and_test_sets#Confusion_in_terminology). 
Splitting the `hindcast-input` (`training-input`) dataset between training and validation is standard way and should be decided carefully.

The `forecast-input` (`test-input`) must not be used as a validation dataset: it must not be used to tune the hyperparameters or make decision about the ML model. 


### Hindcast input

These data are hindcast data. This is used as the input for training the ML models.

The `hindcast-input`(`training-input`) dataset consists of data from three different [models/centers](https://confluence.ecmwf.int/display/S2S/Models):

| center name | model name |
| ----------- | ---------- |
| ecmwf | ecmf | 
| eccc  | cwao |
| ncep  | kwbc |

Use either `origin="ecmf"` (model name) or `origin="ecmwf"` (center name).

This dataset is available as `format`: `grib`, `netcdf`.
  - ECMWF hindcast data
    - `forecast_time`: from 2000/01/02 to 2019/12/31, corresponding to the weekly Thurdays in 2020.
    - `lead_time`: 0 to 46 days
    - `valid_time` (`forecast_time` + `lead_time`): from 2000/01/02 to 2020/02/13
    - availables parameters : `t2m(2t)/siconc(ci)/gh/lsm/msl/q/rsn/sm100/sm20/sp/sst/st100/st20/t/tcc/tcw/tp/ttr/u/v` (differing name in [MARS database](https://confluence.ecmwf.int/display/S2S/Parameters))
  - ECCC hindcast data 
    - `forecast_time`: from 2000/01/02 to 2019/12/31, corresponding to the weekly Thurdays in 2020.
    - `lead_time`: 1 to 32 days
    - `valid_time` (forecast_time + lead_time): from  2000/01/03 to 2020/02/01
    - availables parameters: `t2m(2t)/siconc(ci)/gh/lsm/msl/q/rsn/sp/sst/t/tcc/tcw/tp/ttr/u/v` (differing name in [MARS database](https://confluence.ecmwf.int/display/S2S/Parameters))
    - parameters not available: `sm20/sm100/st20/st100`
  - NCEP hindcast data 
    - `forecast_time` : from 1999/01/07 to 2010/12/30, corresponding to the weekly Thurdays in [2010, see Vitart et al. 2017](https://journals.ametsoc.org/view/journals/bams/98/1/bams-d-16-0017.1.xml), NCEP hindcast dates differ from ECMWF and ECCC hindcasts
    - `lead_time` : 1 to 44 days
    - `valid_time` (`forecast_time` + `lead_time`): from 1999/01/07 to 2011/02/11
    - availables parameters: `t2m(2t)/siconc(ci)/gh/lsm/msl/q/sm100/sm20/sp/sst/st100/st20/t/tcc/tcw/tp/ttr/u/v` (differing name in [MARS database](https://confluence.ecmwf.int/display/S2S/Parameters))
    - parameter not available: `rsn`

```python
hindcast = climetlab.load_dataset('s2s-ai-challenge-training-input', date=[20200102], origin='ecwmf', parameter='tp', format='netcdf').to_xarray()
hindcast.coords
Coordinates:
  * realization    (realization) int64 0 1 2 3 4 5 6 7 8 9 10
  * forecast_time  (forecast_time) datetime64[ns] 2000-01-02 ... 2019-01-02
  * lead_time      (lead_time) timedelta64[ns] 0 days 1 days ... 45 days 46 days
  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0
  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5
    valid_time     (forecast_time, lead_time) datetime64[ns] 2000-01-02 ... 2...
    
# for ncep hindcast provide 2010 date strings
hindcast_ncep = climetlab.load_dataset('s2s-ai-challenge-training-input', date=[20100107], origin='ncep', parameter='tp', format='netcdf').to_xarray()
hindcast_ncep.coords
Coordinates:
  * realization    (realization) int64 0 1 2 3
  * forecast_time  (forecast_time) datetime64[ns] 1999-01-07 ... 2010-01-07
  * lead_time      (lead_time) timedelta64[ns] 1 days 2 days ... 43 days 44 days
  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0
  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5
    valid_time     (forecast_time, lead_time) datetime64[ns] 1999-01-08 ... 2...
```

 List of files :
  [grib](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/grib/index.html),
  [netcdf](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/netcdf/index.html),
  zarr (not available)


### Forecast input

The `forecast-input` (`test-input`) dataset consists also in data from three different models: ECMWF (ecmf), ECCC (cwao), NCEP (eccc), for different dates.
These data are forecast data.
This could be used the input for applying the ML models in order to generate the output which is submitted for the challenge.
  - For all 3 models: 
    - `forecast_time`: from 2020/01/02 to 2020/12/31, weekly every Thurday in 2020.
    - `valid_time` (`forecast_time` + `lead_time`): from 2020/01/02 to 2021/02/xx
    - available parameters (same as for [hindcast input (training input)](#hindcast-input)
  - ECMWF forecast
    - `lead_time`: 0 to 46 days
  - ECCC forecast 
    - `lead_time`: 1 to 32 days
  - NCEP forecast 
    - `lead_time`: 1 to 44 days

```python
forecast = climetlab.load_dataset('s2s-ai-challenge-test-input', date=[20200102], origin='ecmwf', parameter='tp', format='netcdf').to_xarray()
forecast.coords
Coordinates:
  * realization    (realization) int64 0 1 2 3 4 5 6 7 ... 44 45 46 47 48 49 50
  * forecast_time  (forecast_time) datetime64[ns] 2020-01-02
  * lead_time      (lead_time) timedelta64[ns] 0 days 1 days ... 45 days 46 days
  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0
  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5
    valid_time     (forecast_time, lead_time) datetime64[ns] 2020-01-02 ... 2...
```

 List of files :
  [grib](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/test-input/0.3.0/grib/index.html),
  [netcdf](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/test-input/0.3.0/netcdf/index.html),
  zarr (missing)


### Observations (Reference Output)

The `hindcast-like-observations` (`training-output-reference`) dataset matches the `training-input` of the ECMWF and ECCC model.
The `forecast-like-observations` (`test-output-reference`) dataset matches the `test-input` of all three models.

The observations are the ground truth to compare with the ML model output and evaluate them. It consists in observations from instruments of [temperature](http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NCEP/.CPC/.temperature/.daily/) and accumulated total [precipitation](http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NCEP/.CPC/.UNIFIED_PRCP/.GAUGE_BASED/.GLOBAL/.v1p0/.extREALTIME/.rain/). The [NOAA CPC](https://www.cpc.ncep.noaa.gov/) datasets were downloaded from [IRIDL](iridl.ldeo.columbia.edu/). We provide observations in the same dimensions as the forecasts/hindcasts to have an easy match of forecasts/hindcast and ground truth. [See the script for technical details](https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/tree/main/tools/observations).
These observations are the ground truth and do not correspond to a model. The format is always `netcdf`.

```python
hindcast_like_obs = climetlab.load_dataset('s2s-ai-challenge-training-output-reference', date=[20200102], parameter='tp').to_xarray()  # origin and format not accepted
hindcast_like_obs.coords
Coordinates:
  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0
  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5
  * forecast_time  (forecast_time) datetime64[ns] 2000-01-02 ... 2019-01-02
  * lead_time      (lead_time) timedelta64[ns] 0 days 1 days ... 45 days 46 days
    valid_time     (lead_time, forecast_time) datetime64[ns] 2000-01-02 ... 2...

forecast_like_obs = climetlab.load_dataset('s2s-ai-challenge-test-output-reference', date=[20200102], parameter='tp').to_xarray()  # origin and format not accepted
forecast_like_obs.coords
Coordinates:
  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0
  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5
  * forecast_time  (forecast_time) datetime64[ns] 2020-01-02
  * lead_time      (lead_time) timedelta64[ns] 0 days 1 days ... 45 days 46 days
    valid_time     (lead_time, forecast_time) datetime64[ns] 2020-01-02 ... 2...
```

In case you want to train on the NCEP hindcast, which has `forecast_time`s from 1999 to 2010, please use download observations with a time dimension `s2s-ai-challenge-observations` and use [`climetlab_s2s_ai_challenge.extra.forecast_like_observations`](https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/climetlab_s2s_ai_challenge/extra.py#L40) to match observations to the corresponding `valid_time`s of the forecast/hindcast.

```python
obs_time = climetlab.load_dataset('s2s-ai-challenge-observations', parameter=['pr', 't2m']).to_xarray()
# equivalent
obs_lead_time_forecast_time = climetlab.load_dataset('s2s-ai-challenge-observations', parameter=['pr', 't2m']).to_xarray(like=hindcast_ncep)
obs_lead_time_forecast_time = climetlab_s2s_ai_challenge.extra.forecast_like_observations(hindcast_ncep, obs_time)
obs_lead_time_forecast_time.coords
<xarray.Dataset>
Dimensions:        (forecast_time: 12, latitude: 121, lead_time: 44, longitude: 240)
Coordinates:
    valid_time     (forecast_time, lead_time) datetime64[ns] 1999-01-08 ... 2...
  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0
  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5
  * forecast_time  (forecast_time) datetime64[ns] 1999-01-07 ... 2010-01-07
  * lead_time      (lead_time) timedelta64[ns] 1 days 2 days ... 43 days 44 days
```

This function can be used for all initialized hindcasts and forecasts from the SubX and S2S projects. Beware of the different starting dates of 2020 forecasts when using them for the `s2s-ai-challenge`. Observations from NOAA CPC are [regridded conservatively](https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/blob/4829222db2de02d4255cef8dfd48c986365cc8be/tools/observations/build_dataset_observations.py#L75) with [`xesmf`](https://pangeo-xesmf.readthedocs.io/en/latest/) to the S2S 1.5 deg grid. The original 0.5 degree spatial resolution raw observations can be obtained via `grid='720x360'`.


Generally speaking, only data available when the forecast is issued can be used by the ML models to perform their forecast:

__Rule: Observed data beyond the forecast date should not be used for prediction, for instance a forecast starting on 2020/07/01 should not use observed data beyond 2020/07/01).__

For all rules, see the [challenge website](https://s2s-ai-challenge.github.io/#rules).

See also the general rules of the challenge [here](https://s2s-ai-challenge.github.io/#rules).

Dates in the observation dataset are from 2000/01/01 to 2021/02/15.

The observations dataset have been build from real instrument observations.

- The `hindcast-like-observations` (`training-output-reference`) dataset :
 - Available from 2000/01/01 to 2019/12/31, weekly every 7 days (every Thurday)
 - Observation data before 2019/12/31 can be used for training (as the truth to evaluate and optimize the ML models or tweak hyper parameters using train/valid split or cross-validation).
- The `forecast-like-observations` (`test-output-reference`) dataset.
 - Available from 2020/01/01 to 2021/02/20 , weekly every 7 days (every Thurday)
 - The test data must **not** be used during training. In theory, these data should not be disclosed during the challenge, but the nature of the data make is possible to access it from other sources. That is the reason why the code used for training model must be submitted along with the prediction (as a jupyter notebook) and the top ranked proposition will be reviewed by the organizing board. 

![train_validation_split](https://user-images.githubusercontent.com/8441217/114999589-e5f29f80-9e99-11eb-90e3-8a4a3e9545d5.png)

During forecast phase (i.e. the evaluation phase using the forecast-input dataset), 2020 observation data is used. Rule 1 still stands: Observed data beyond the forecast start date should not be used for prediction.

### Forecast Benchmark (Benchmark output)

The `forecast-benchmark` (`test-output-benchmark`) dataset is a probabilistic re-calibrated ECMWF forecast with categories `below normal`, `near normal`, `above normal`. The calibration has been performed by using the tercile boundaries from the model climatology rather than from observations.

The benchmark data is available as follows:
  - `forecast_time`: from 2020/01/02 to 2020/12/31, weekly every 7 days (every Thurday).
  - `lead_time`: 14 days and 28 days, where this day represents the first day of the biweekly aggregate
  - `valid_time` (`forecast_time` + `lead_time`): from 2020/01/01 to 2021/01/29
  - `category`: `'below normal'`, `'near normal'`, `'above normal'`

```python
bench = climetlab.load_dataset('s2s-ai-challenge-training-output-benchmark', parameter='tp').to_xarray()  # origin, date and format not accepted
bench.coords
Coordinates:
  * category       (category) object 'below normal' 'near normal' 'above normal'
  * forecast_time  (forecast_time) datetime64[ns] 2000-01-02 ... 2019-12-31
  * lead_time      (lead_time) timedelta64[ns] 14 days 28 days
  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0
  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5
    valid_time     (forecast_time, lead_time) datetime64[ns] 2000-01-16 ... 2...

bench = climetlab.load_dataset('s2s-ai-challenge-test-output-benchmark', parameter='tp').to_xarray()  # origin, date and format not accepted
bench.coords
Coordinates:
  * category       (category) object 'below normal' 'near normal' 'above normal'
  * forecast_time  (forecast_time) datetime64[ns] 2020-01-02 ... 2020-12-31
  * lead_time      (lead_time) timedelta64[ns] 14 days 28 days
  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0
  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5
    valid_time     (forecast_time, lead_time) datetime64[ns] 2020-01-16 ... 2...
```

### Observations

For other initialized forecasts, you can download `observations` for parameters `t2m` and `pr` with a `time` dimension corresponding to `valid_time`. These are then used to create observations formatted like initialized forecasts/hindcasts locally. Observations are available from 1999 to 2021. See [parameter](#parameter) for description.

[`climetlab_s2s_ai_challenge.extra.forecast_like_observations`](https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/climetlab_s2s_ai_challenge/extra.py#L40) matches observations to the corresponding `valid_time`s of the forecast/hindcast.

```python
forecast = climetlab.load_dataset('s2s-ai-challenge-training-input',
        date=20100107, origin='ncep', parameter='tp',
        format='netcdf').to_xarray()

obs_lead_time_forecast_time = climetlab.load_dataset('s2s-ai-challenge-observations', parameter=['pr', 't2m']).to_xarray(like=forecast)
# equivalent
obs_time = climetlab.load_dataset('s2s-ai-challenge-observations', parameter=['pr', 't2m']).to_xarray()
obs_time.coords
Coordinates:
  * time       (time) datetime64[ns] 1999-01-01 1999-01-02 ... 2021-04-29
  * latitude   (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0
  * longitude  (longitude) float64 0.0 1.5 3.0 4.5 ... 354.0 355.5 357.0 358.5

obs_lead_time_forecast_time = climetlab_s2s_ai_challenge.extra.forecast_like_observations(forecast, obs_time)
obs_lead_time_forecast_time
<xarray.Dataset>
Dimensions:        (forecast_time: 12, latitude: 121, lead_time: 44, longitude: 240)
Coordinates:
    valid_time     (forecast_time, lead_time) datetime64[ns] 1999-01-08 ... 2...
  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0
  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5
  * forecast_time  (forecast_time) datetime64[ns] 1999-01-07 ... 2010-01-07
  * lead_time      (lead_time) timedelta64[ns] 1 days 2 days ... 43 days 44 days
Data variables:
    t2m            (forecast_time, lead_time, latitude, longitude) float32 ...
    tp             (forecast_time, lead_time, latitude, longitude) float32 na...
Attributes:
    script:   climetlab_s2s_ai_challenge.extra.forecast_like_observations
```

# Data download (GRIB or NetCDF)

The URLs to download the data are constructed according to the following patterns: 

*For input datasets*, the pattern is https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/{datasetname}/0.3.0/{format}/{origin}-{fctype}-{parameter}-YYYYMMDD.nc

*For observations datasets (reference output)*, the pattern is https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/{datasetname}/{parameter}-YYYYMMDD.nc

*For benchmark datasets*, the pattern will be similar to https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-output-benchmark/{parameter}.nc


- {datasetname} : In the URLs the dataset name must follow the ML naming (`training-input` or `training-output-reference` or `training-output-benchmark`).
- {format} is `netcdf`. Training output is also available as GRIB file,  using `format='grib'` and replacing `".nc"` by `".grib"`
- {fctype}: `hindcast` for training or `forecast` for test
- {parameter} is `t2m` for [surface temperature at 2m](https://confluence.ecmwf.int/display/S2S/S2S+Surface+Air+Temperature), `tp` for [total precipitation](https://confluence.ecmwf.int/display/S2S/S2S+Total+Precipitation)
- {origin} : `ecmwf` or `eccc` or `ncep`
- {weeks} from [`"34"`, `"56"`, `["34", "56"]`] only for `benchmark`
- {grid} from [`"240x121"` (default), `"720x360"`] only for `observations`
- `YYYYMMDD` is the date of the 2020 forecast for `test-input`/`forecast-input`. The same dates are required for the on-the-fly `training-input`/`hindcast-input` but return the multi-year hindcast for that `MMDD`.

The list of files for the `training-input` dataset can be found at
  - GRIB: [https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/grib/index.html](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/grib/index.html),
 - NetCDF: [https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/netcdf/index.html](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/netcdf/index.html),

The list of files for the `training-output-benchmark` dataset can be found at [https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-output-benchmark/index.html](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-output-benchmark/index.html) (NetCDF only) 

The list of files for the `training-output-reference` dataset can be found at [https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-output-reference/index.html](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-output-reference/index.html) (NetCDF only) 


Example to retrieve the file with wget :

``` wget https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/grib/ncep-hindcast-q-20101014.grib ``` (132.8M )

### Zarr format (experimental).
 
The zarr storage location include all the reference data. The zarr urls are **not** designed to be open in a browser (see [zarr](https://zarr.readthedocs.io/en/stable)):
While accessing the zarr storage without climetlab may be possible, we recommend using climetlab with the appropriate plugin (climetlab-s2s-ai-challenge)

Zarr urls are not available.
  

## Using climetlab to access the data (supports grib, netcdf and zarr)

The climetlab python package allows easy access to the data with a few lines of code such as:

```
!pip install climetlab climetlab_s2s_ai_challenge
import climetlab as cml
cml.load_dataset("s2s-ai-challenge-training-input",
                         origin='ecmwf',
                         date=[20200102,20200109],
                         # optional : format='grib'
                         parameter='tp').to_xarray()
cml.load_dataset("s2s-ai-challenge-training-output-reference",
                         date=[20200102,20200109],
                         parameter='tp').to_xarray()
```

See the demo notebooks here: https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/notebooks.

Accessing the training data :
- Netcdf [nbviewer](https://nbviewer.jupyter.org/github/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/notebooks/demo_netcdf.ipynb) [colab](https://colab.research.google.com/github/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/notebooks/demo_netcdf.ipynb)
- Grib [nbviewer](https://nbviewer.jupyter.org/github/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/notebooks/demo_grib.ipynb) [colab](https://colab.research.google.com/github/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/notebooks/demo_grib.ipynb)


Getting the observation (reference output) dataset see the [demo_observations notebook](https://nbviewer.jupyter.org/github/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/notebooks/demo_observations.ipynb).

Getting the benchmark dataset see the [demo_benchmark notebook](https://nbviewer.jupyter.org/github/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/notebooks/demo_benchmark.ipynb).
            

Raw data

            {
    "_id": null,
    "home_page": "https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge",
    "name": "climetlab-s2s-ai-challenge",
    "maintainer": "",
    "docs_url": null,
    "requires_python": "",
    "maintainer_email": "",
    "keywords": "meteorology",
    "author": "European Centre for Medium-Range Weather Forecasts (ECMWF)",
    "author_email": "software.support@ecmwf.int",
    "download_url": "https://files.pythonhosted.org/packages/c8/eb/288466bbf139d84222f920a5f8ef2db271089964679eb85dd4b727a1133f/climetlab-s2s-ai-challenge-0.9.0.tar.gz",
    "platform": null,
    "description": "[![Check and publish Python Package](https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/actions/workflows/check-and-publish.yml/badge.svg)](https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/actions/workflows/check-and-publish.yml) \n[![PyPI version fury.io](https://badge.fury.io/py/climetlab-s2s-ai-challenge.svg)](https://pypi.python.org/pypi/climetlab-s2s-ai-challenge/)\n[![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/ecmwf-lab/climetlab-s2s-ai-challenge/main?urlpath=lab)\n\n# S2S AI challenge Datasets\n\nThis is a `climetlab` plugin for Sub-seasonal to Seasonal (S2S) Artificial Intelligence Challenge: https://s2s-ai-challenge.github.io/.\n\nIn this README is a description of how to get the data for the S2S AI challenge. Here is a more general [description of the S2S data](https://confluence.ecmwf.int/display/S2S/Description). The data used for the S2S AI challenge is a subset of the S2S data library. More detail can be found at https://confluence.ecmwf.int/display/S2S  and https://confluence.ecmwf.int/display/S2S/Parameters.\n\nThere are several ways to use the datasets. Either by direct download (`wget`, `curl`, `browser`) for [`GRIB`](https://en.wikipedia.org/wiki/GRIB) and [NetCDF](https://en.wikipedia.org/wiki/NetCDF) formats; or using the `climetlab` python package with this addon, for `GRIB` and `NetCDF` and `zarr` formats. [`zarr`](https://zarr.readthedocs.io/en/stable/) is a cloud-friendly experimental data format and supports dowloading only the part of the data that is required. It has been designed to work better than classical format on a cloud environment (experimental).\n\n# Installation\n\n`pip install -U climetlab climetlab_s2s_ai_challenge`\n\n\n# API\n\n```python\nimport climetlab\n```\n\nUse `climetlab.load_dataset('s2s-ai-challenge-{datasetname}')` with the following keywords:\n\n- `datasetname`: name of the dataset, see [dataset description](#datasets-description)\n- `parameter`: [variable](#parameter), see [hindcast input for the different models](#hindcast-input)\n- `origin`: name of the model [`ecmwf`, `eccc`, `ncep`] or modelling center [`ecmf`, `cwao`, `kwbc`]. Provide `origin` only for `training/test-input`/`hindcast/forecast-input`.\n- `date`: `YYYYMMDD` is the date of the 2020 forecast for `test-input`/`forecast-input`. The same dates are required for the on-the-fly `training-input`/`hindcast-input` but returns the multi-year hindcast for the given `MMDD` date. Please provide `int`, `str`, `np.datetime` or list of the former in format `YYYYMMDDD`. Providing no `date` keyword downloads all dates.\n- `format`: data format, choose from [`netcdf` (always available), `grb` (only for `input-*`), `zarr` (experimental)]\n\n## Coordinates\n\nOverview of the time-related coordinates.\n\n| `name` | CF convention `standard_name` | description | comment |\n| --- | --- | --- | --- |\n| `forecast_time` | forecast_reference_time | The forecast reference time in NWP is the \"data time\", the time of the analysis from which the forecast was made. It is not the time for which the forecast is valid. | |\n| `lead_time` | forecast_period | Forecast period is the time interval between the forecast reference time and the validity time. |  |\n| `valid_time` | time | time for which the forecast is valid | `forecast_time` + `lead_time` |\n\nAll datasets are on a global 1.5 degree grid.\n\n\n## Parameter\n\n`parameter` describes the variable to download. The most important two variables for the `s2s-ai-challenge` are the two target variables `t2m` and `tp`:\n\n| parameter | long_name | standard_name | unit | description & aggregation type | week 3-4 aggregation | week 5-6 aggregation | link to source |\n| -------- | --------- | --- | --- | --- | --- | --- | --- |\n| `t2m`    | 2m temperature | air_temperature | K | Temperature at 2m height averaged for the date given | average [day 14, day 27] | average [day 28, day 41] | [model](https://confluence.ecmwf.int/display/S2S/S2S+Surface+Air+Temperature), [observations](http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NCEP/.CPC/.temperature/.daily/)|\n| `tp`     | total precipitation | precipitation_amount | kg m-2 | Total precipitation accumulated from `forecast_time` until including `valid_time`, e.g. `lead_time = 1 days` accumulates precipitation_flux `pr` from 6-hourly steps 0,6,12,18 at date `forecast_time` | day 28 minus day 14 | day 42 minus day 28 | [model](https://confluence.ecmwf.int/display/S2S/S2S+Total+Precipitation) | \n| `pr`     | precipitation flux | precipitation_flux | kg m-2 | Precipitation accumulated for the date given | use `tp` | use `tp` | [observations](http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NCEP/.CPC/.UNIFIED_PRCP/.GAUGE_BASED/.GLOBAL/.v1p0/.extREALTIME/.rain) |\n\nGiven the different nature of the parameters, with `forecast_time` Jan 2nd 2020, `tp` `lead_time=1 days` and `t2m` `lead_time=0 days` describe both the weather of Jan 2nd 2020, as `tp` is aggregated `pr` from Jan 2nd 00:00 to Jan 3rd 00:00 and `t2m` as the average of Jan 2nd. Furthermore `tp` is aggregated since `forecast_time`, i.e. `tp` `lead_time=5 days` is `pr` aggregated from Jan 2nd 00:00 to Jan 7th 00:00.\n\nFor the remaining variable descriptions, see [ECWMF S2S description](https://confluence.ecmwf.int/display/S2S/Parameters).\n\n\n## Datasets description\n\nThere are four datasets provided for this challenge. As we are aiming at bringing together the two communities of Machine Learning and Weather Prediction, they have been aliases to use both two points of views:\n\n| ML                          | NWP                          |                                                        |\n| --------------------------- | ---------------------------- | ------------------------------------------------------ |\n| `training-input`            | `hindcast-input`             | Training  dataset (input for training the ML models)   |\n| `training-output-reference` | `hindcast-like-observations` | Training dataset (output for training the ML models)   |\n| `training-output-benchmark` | `hindcast-benchmark`         | Benchmark output (on the training dataset)             |\n| `test-input`                | `forecast-input`             | Test dataset (DO NOT use for training)                 |\n| `test-output-reference`     | `forecast-like-observations` | Test dataset (DO NOT use)                              |\n| `test-output-benchmark`     | `forecast-benchmark`         | Benchmark output (on the test dataset)                 |\n| `observations`              | `observations`               | Observations with `time` dimension                     |\n\n\n**Overfitting** is always an potential issue when using ML algorithms. To address this, the data is usually split into three datasets : \n[training](https://en.wikipedia.org/wiki/Training,_validation,_and_test_sets#Training_dataset), \n[validation](https://en.wikipedia.org/wiki/Training,_validation,_and_test_sets#Validation_dataset) \nand [test](https://en.wikipedia.org/wiki/Training,_validation,_and_test_sets#Test_dataset). \nThis terminology has lead to [some confusion in the past](https://en.wikipedia.org/wiki/Training,_validation,_and_test_sets#Confusion_in_terminology). \nSplitting the `hindcast-input` (`training-input`) dataset between training and validation is standard way and should be decided carefully.\n\nThe `forecast-input` (`test-input`) must not be used as a validation dataset: it must not be used to tune the hyperparameters or make decision about the ML model. \n\n\n### Hindcast input\n\nThese data are hindcast data. This is used as the input for training the ML models.\n\nThe `hindcast-input`(`training-input`) dataset consists of data from three different [models/centers](https://confluence.ecmwf.int/display/S2S/Models):\n\n| center name | model name |\n| ----------- | ---------- |\n| ecmwf | ecmf | \n| eccc  | cwao |\n| ncep  | kwbc |\n\nUse either `origin=\"ecmf\"` (model name) or `origin=\"ecmwf\"` (center name).\n\nThis dataset is available as `format`: `grib`, `netcdf`.\n  - ECMWF hindcast data\n    - `forecast_time`: from 2000/01/02 to 2019/12/31, corresponding to the weekly Thurdays in 2020.\n    - `lead_time`: 0 to 46 days\n    - `valid_time` (`forecast_time` + `lead_time`): from 2000/01/02 to 2020/02/13\n    - availables parameters : `t2m(2t)/siconc(ci)/gh/lsm/msl/q/rsn/sm100/sm20/sp/sst/st100/st20/t/tcc/tcw/tp/ttr/u/v` (differing name in [MARS database](https://confluence.ecmwf.int/display/S2S/Parameters))\n  - ECCC hindcast data \n    - `forecast_time`: from 2000/01/02 to 2019/12/31, corresponding to the weekly Thurdays in 2020.\n    - `lead_time`: 1 to 32 days\n    - `valid_time` (forecast_time + lead_time): from  2000/01/03 to 2020/02/01\n    - availables parameters: `t2m(2t)/siconc(ci)/gh/lsm/msl/q/rsn/sp/sst/t/tcc/tcw/tp/ttr/u/v` (differing name in [MARS database](https://confluence.ecmwf.int/display/S2S/Parameters))\n    - parameters not available: `sm20/sm100/st20/st100`\n  - NCEP hindcast data \n    - `forecast_time` : from 1999/01/07 to 2010/12/30, corresponding to the weekly Thurdays in [2010, see Vitart et al. 2017](https://journals.ametsoc.org/view/journals/bams/98/1/bams-d-16-0017.1.xml), NCEP hindcast dates differ from ECMWF and ECCC hindcasts\n    - `lead_time` : 1 to 44 days\n    - `valid_time` (`forecast_time` + `lead_time`): from 1999/01/07 to 2011/02/11\n    - availables parameters: `t2m(2t)/siconc(ci)/gh/lsm/msl/q/sm100/sm20/sp/sst/st100/st20/t/tcc/tcw/tp/ttr/u/v` (differing name in [MARS database](https://confluence.ecmwf.int/display/S2S/Parameters))\n    - parameter not available: `rsn`\n\n```python\nhindcast = climetlab.load_dataset('s2s-ai-challenge-training-input', date=[20200102], origin='ecwmf', parameter='tp', format='netcdf').to_xarray()\nhindcast.coords\nCoordinates:\n  * realization    (realization) int64 0 1 2 3 4 5 6 7 8 9 10\n  * forecast_time  (forecast_time) datetime64[ns] 2000-01-02 ... 2019-01-02\n  * lead_time      (lead_time) timedelta64[ns] 0 days 1 days ... 45 days 46 days\n  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0\n  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5\n    valid_time     (forecast_time, lead_time) datetime64[ns] 2000-01-02 ... 2...\n    \n# for ncep hindcast provide 2010 date strings\nhindcast_ncep = climetlab.load_dataset('s2s-ai-challenge-training-input', date=[20100107], origin='ncep', parameter='tp', format='netcdf').to_xarray()\nhindcast_ncep.coords\nCoordinates:\n  * realization    (realization) int64 0 1 2 3\n  * forecast_time  (forecast_time) datetime64[ns] 1999-01-07 ... 2010-01-07\n  * lead_time      (lead_time) timedelta64[ns] 1 days 2 days ... 43 days 44 days\n  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0\n  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5\n    valid_time     (forecast_time, lead_time) datetime64[ns] 1999-01-08 ... 2...\n```\n\n List of files :\n  [grib](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/grib/index.html),\n  [netcdf](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/netcdf/index.html),\n  zarr (not available)\n\n\n### Forecast input\n\nThe `forecast-input` (`test-input`) dataset consists also in data from three different models: ECMWF (ecmf), ECCC (cwao), NCEP (eccc), for different dates.\nThese data are forecast data.\nThis could be used the input for applying the ML models in order to generate the output which is submitted for the challenge.\n  - For all 3 models: \n    - `forecast_time`: from 2020/01/02 to 2020/12/31, weekly every Thurday in 2020.\n    - `valid_time` (`forecast_time` + `lead_time`): from 2020/01/02 to 2021/02/xx\n    - available parameters (same as for [hindcast input (training input)](#hindcast-input)\n  - ECMWF forecast\n    - `lead_time`: 0 to 46 days\n  - ECCC forecast \n    - `lead_time`: 1 to 32 days\n  - NCEP forecast \n    - `lead_time`: 1 to 44 days\n\n```python\nforecast = climetlab.load_dataset('s2s-ai-challenge-test-input', date=[20200102], origin='ecmwf', parameter='tp', format='netcdf').to_xarray()\nforecast.coords\nCoordinates:\n  * realization    (realization) int64 0 1 2 3 4 5 6 7 ... 44 45 46 47 48 49 50\n  * forecast_time  (forecast_time) datetime64[ns] 2020-01-02\n  * lead_time      (lead_time) timedelta64[ns] 0 days 1 days ... 45 days 46 days\n  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0\n  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5\n    valid_time     (forecast_time, lead_time) datetime64[ns] 2020-01-02 ... 2...\n```\n\n List of files :\n  [grib](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/test-input/0.3.0/grib/index.html),\n  [netcdf](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/test-input/0.3.0/netcdf/index.html),\n  zarr (missing)\n\n\n### Observations (Reference Output)\n\nThe `hindcast-like-observations` (`training-output-reference`) dataset matches the `training-input` of the ECMWF and ECCC model.\nThe `forecast-like-observations` (`test-output-reference`) dataset matches the `test-input` of all three models.\n\nThe observations are the ground truth to compare with the ML model output and evaluate them. It consists in observations from instruments of [temperature](http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NCEP/.CPC/.temperature/.daily/) and accumulated total [precipitation](http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NCEP/.CPC/.UNIFIED_PRCP/.GAUGE_BASED/.GLOBAL/.v1p0/.extREALTIME/.rain/). The [NOAA CPC](https://www.cpc.ncep.noaa.gov/) datasets were downloaded from [IRIDL](iridl.ldeo.columbia.edu/). We provide observations in the same dimensions as the forecasts/hindcasts to have an easy match of forecasts/hindcast and ground truth. [See the script for technical details](https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/tree/main/tools/observations).\nThese observations are the ground truth and do not correspond to a model. The format is always `netcdf`.\n\n```python\nhindcast_like_obs = climetlab.load_dataset('s2s-ai-challenge-training-output-reference', date=[20200102], parameter='tp').to_xarray()  # origin and format not accepted\nhindcast_like_obs.coords\nCoordinates:\n  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0\n  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5\n  * forecast_time  (forecast_time) datetime64[ns] 2000-01-02 ... 2019-01-02\n  * lead_time      (lead_time) timedelta64[ns] 0 days 1 days ... 45 days 46 days\n    valid_time     (lead_time, forecast_time) datetime64[ns] 2000-01-02 ... 2...\n\nforecast_like_obs = climetlab.load_dataset('s2s-ai-challenge-test-output-reference', date=[20200102], parameter='tp').to_xarray()  # origin and format not accepted\nforecast_like_obs.coords\nCoordinates:\n  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0\n  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5\n  * forecast_time  (forecast_time) datetime64[ns] 2020-01-02\n  * lead_time      (lead_time) timedelta64[ns] 0 days 1 days ... 45 days 46 days\n    valid_time     (lead_time, forecast_time) datetime64[ns] 2020-01-02 ... 2...\n```\n\nIn case you want to train on the NCEP hindcast, which has `forecast_time`s from 1999 to 2010, please use download observations with a time dimension `s2s-ai-challenge-observations` and use [`climetlab_s2s_ai_challenge.extra.forecast_like_observations`](https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/climetlab_s2s_ai_challenge/extra.py#L40) to match observations to the corresponding `valid_time`s of the forecast/hindcast.\n\n```python\nobs_time = climetlab.load_dataset('s2s-ai-challenge-observations', parameter=['pr', 't2m']).to_xarray()\n# equivalent\nobs_lead_time_forecast_time = climetlab.load_dataset('s2s-ai-challenge-observations', parameter=['pr', 't2m']).to_xarray(like=hindcast_ncep)\nobs_lead_time_forecast_time = climetlab_s2s_ai_challenge.extra.forecast_like_observations(hindcast_ncep, obs_time)\nobs_lead_time_forecast_time.coords\n<xarray.Dataset>\nDimensions:        (forecast_time: 12, latitude: 121, lead_time: 44, longitude: 240)\nCoordinates:\n    valid_time     (forecast_time, lead_time) datetime64[ns] 1999-01-08 ... 2...\n  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0\n  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5\n  * forecast_time  (forecast_time) datetime64[ns] 1999-01-07 ... 2010-01-07\n  * lead_time      (lead_time) timedelta64[ns] 1 days 2 days ... 43 days 44 days\n```\n\nThis function can be used for all initialized hindcasts and forecasts from the SubX and S2S projects. Beware of the different starting dates of 2020 forecasts when using them for the `s2s-ai-challenge`. Observations from NOAA CPC are [regridded conservatively](https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/blob/4829222db2de02d4255cef8dfd48c986365cc8be/tools/observations/build_dataset_observations.py#L75) with [`xesmf`](https://pangeo-xesmf.readthedocs.io/en/latest/) to the S2S 1.5 deg grid. The original 0.5 degree spatial resolution raw observations can be obtained via `grid='720x360'`.\n\n\nGenerally speaking, only data available when the forecast is issued can be used by the ML models to perform their forecast:\n\n__Rule: Observed data beyond the forecast date should not be used for prediction, for instance a forecast starting on 2020/07/01 should not use observed data beyond 2020/07/01).__\n\nFor all rules, see the [challenge website](https://s2s-ai-challenge.github.io/#rules).\n\nSee also the general rules of the challenge [here](https://s2s-ai-challenge.github.io/#rules).\n\nDates in the observation dataset are from 2000/01/01 to 2021/02/15.\n\nThe observations dataset have been build from real instrument observations.\n\n- The `hindcast-like-observations` (`training-output-reference`) dataset :\n - Available from 2000/01/01 to 2019/12/31, weekly every 7 days (every Thurday)\n - Observation data before 2019/12/31 can be used for training (as the truth to evaluate and optimize the ML models or tweak hyper parameters using train/valid split or cross-validation).\n- The `forecast-like-observations` (`test-output-reference`) dataset.\n - Available from 2020/01/01 to 2021/02/20 , weekly every 7 days (every Thurday)\n - The test data must **not** be used during training. In theory, these data should not be disclosed during the challenge, but the nature of the data make is possible to access it from other sources. That is the reason why the code used for training model must be submitted along with the prediction (as a jupyter notebook) and the top ranked proposition will be reviewed by the organizing board. \n\n![train_validation_split](https://user-images.githubusercontent.com/8441217/114999589-e5f29f80-9e99-11eb-90e3-8a4a3e9545d5.png)\n\nDuring forecast phase (i.e. the evaluation phase using the forecast-input dataset), 2020 observation data is used. Rule 1 still stands: Observed data beyond the forecast start date should not be used for prediction.\n\n### Forecast Benchmark (Benchmark output)\n\nThe `forecast-benchmark` (`test-output-benchmark`) dataset is a probabilistic re-calibrated ECMWF forecast with categories `below normal`, `near normal`, `above normal`. The calibration has been performed by using the tercile boundaries from the model climatology rather than from observations.\n\nThe benchmark data is available as follows:\n  - `forecast_time`: from 2020/01/02 to 2020/12/31, weekly every 7 days (every Thurday).\n  - `lead_time`: 14 days and 28 days, where this day represents the first day of the biweekly aggregate\n  - `valid_time` (`forecast_time` + `lead_time`): from 2020/01/01 to 2021/01/29\n  - `category`: `'below normal'`, `'near normal'`, `'above normal'`\n\n```python\nbench = climetlab.load_dataset('s2s-ai-challenge-training-output-benchmark', parameter='tp').to_xarray()  # origin, date and format not accepted\nbench.coords\nCoordinates:\n  * category       (category) object 'below normal' 'near normal' 'above normal'\n  * forecast_time  (forecast_time) datetime64[ns] 2000-01-02 ... 2019-12-31\n  * lead_time      (lead_time) timedelta64[ns] 14 days 28 days\n  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0\n  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5\n    valid_time     (forecast_time, lead_time) datetime64[ns] 2000-01-16 ... 2...\n\nbench = climetlab.load_dataset('s2s-ai-challenge-test-output-benchmark', parameter='tp').to_xarray()  # origin, date and format not accepted\nbench.coords\nCoordinates:\n  * category       (category) object 'below normal' 'near normal' 'above normal'\n  * forecast_time  (forecast_time) datetime64[ns] 2020-01-02 ... 2020-12-31\n  * lead_time      (lead_time) timedelta64[ns] 14 days 28 days\n  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0\n  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5\n    valid_time     (forecast_time, lead_time) datetime64[ns] 2020-01-16 ... 2...\n```\n\n### Observations\n\nFor other initialized forecasts, you can download `observations` for parameters `t2m` and `pr` with a `time` dimension corresponding to `valid_time`. These are then used to create observations formatted like initialized forecasts/hindcasts locally. Observations are available from 1999 to 2021. See [parameter](#parameter) for description.\n\n[`climetlab_s2s_ai_challenge.extra.forecast_like_observations`](https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/climetlab_s2s_ai_challenge/extra.py#L40) matches observations to the corresponding `valid_time`s of the forecast/hindcast.\n\n```python\nforecast = climetlab.load_dataset('s2s-ai-challenge-training-input',\n        date=20100107, origin='ncep', parameter='tp',\n        format='netcdf').to_xarray()\n\nobs_lead_time_forecast_time = climetlab.load_dataset('s2s-ai-challenge-observations', parameter=['pr', 't2m']).to_xarray(like=forecast)\n# equivalent\nobs_time = climetlab.load_dataset('s2s-ai-challenge-observations', parameter=['pr', 't2m']).to_xarray()\nobs_time.coords\nCoordinates:\n  * time       (time) datetime64[ns] 1999-01-01 1999-01-02 ... 2021-04-29\n  * latitude   (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0\n  * longitude  (longitude) float64 0.0 1.5 3.0 4.5 ... 354.0 355.5 357.0 358.5\n\nobs_lead_time_forecast_time = climetlab_s2s_ai_challenge.extra.forecast_like_observations(forecast, obs_time)\nobs_lead_time_forecast_time\n<xarray.Dataset>\nDimensions:        (forecast_time: 12, latitude: 121, lead_time: 44, longitude: 240)\nCoordinates:\n    valid_time     (forecast_time, lead_time) datetime64[ns] 1999-01-08 ... 2...\n  * latitude       (latitude) float64 90.0 88.5 87.0 85.5 ... -87.0 -88.5 -90.0\n  * longitude      (longitude) float64 0.0 1.5 3.0 4.5 ... 355.5 357.0 358.5\n  * forecast_time  (forecast_time) datetime64[ns] 1999-01-07 ... 2010-01-07\n  * lead_time      (lead_time) timedelta64[ns] 1 days 2 days ... 43 days 44 days\nData variables:\n    t2m            (forecast_time, lead_time, latitude, longitude) float32 ...\n    tp             (forecast_time, lead_time, latitude, longitude) float32 na...\nAttributes:\n    script:   climetlab_s2s_ai_challenge.extra.forecast_like_observations\n```\n\n# Data download (GRIB or NetCDF)\n\nThe URLs to download the data are constructed according to the following patterns: \n\n*For input datasets*, the pattern is https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/{datasetname}/0.3.0/{format}/{origin}-{fctype}-{parameter}-YYYYMMDD.nc\n\n*For observations datasets (reference output)*, the pattern is https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/{datasetname}/{parameter}-YYYYMMDD.nc\n\n*For benchmark datasets*, the pattern will be similar to https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-output-benchmark/{parameter}.nc\n\n\n- {datasetname} : In the URLs the dataset name must follow the ML naming (`training-input` or `training-output-reference` or `training-output-benchmark`).\n- {format} is `netcdf`. Training output is also available as GRIB file,  using `format='grib'` and replacing `\".nc\"` by `\".grib\"`\n- {fctype}: `hindcast` for training or `forecast` for test\n- {parameter} is `t2m` for [surface temperature at 2m](https://confluence.ecmwf.int/display/S2S/S2S+Surface+Air+Temperature), `tp` for [total precipitation](https://confluence.ecmwf.int/display/S2S/S2S+Total+Precipitation)\n- {origin} : `ecmwf` or `eccc` or `ncep`\n- {weeks} from [`\"34\"`, `\"56\"`, `[\"34\", \"56\"]`] only for `benchmark`\n- {grid} from [`\"240x121\"` (default), `\"720x360\"`] only for `observations`\n- `YYYYMMDD` is the date of the 2020 forecast for `test-input`/`forecast-input`. The same dates are required for the on-the-fly `training-input`/`hindcast-input` but return the multi-year hindcast for that `MMDD`.\n\nThe list of files for the `training-input` dataset can be found at\n  - GRIB: [https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/grib/index.html](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/grib/index.html),\n - NetCDF: [https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/netcdf/index.html](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/netcdf/index.html),\n\nThe list of files for the `training-output-benchmark` dataset can be found at [https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-output-benchmark/index.html](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-output-benchmark/index.html) (NetCDF only) \n\nThe list of files for the `training-output-reference` dataset can be found at [https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-output-reference/index.html](https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-output-reference/index.html) (NetCDF only) \n\n\nExample to retrieve the file with wget :\n\n``` wget https://object-store.os-api.cci1.ecmwf.int/s2s-ai-challenge/data/training-input/0.3.0/grib/ncep-hindcast-q-20101014.grib ``` (132.8M )\n\n### Zarr format (experimental).\n \nThe zarr storage location include all the reference data. The zarr urls are **not** designed to be open in a browser (see [zarr](https://zarr.readthedocs.io/en/stable)):\nWhile accessing the zarr storage without climetlab may be possible, we recommend using climetlab with the appropriate plugin (climetlab-s2s-ai-challenge)\n\nZarr urls are not available.\n  \n\n## Using climetlab to access the data (supports grib, netcdf and zarr)\n\nThe climetlab python package allows easy access to the data with a few lines of code such as:\n\n```\n!pip install climetlab climetlab_s2s_ai_challenge\nimport climetlab as cml\ncml.load_dataset(\"s2s-ai-challenge-training-input\",\n                         origin='ecmwf',\n                         date=[20200102,20200109],\n                         # optional : format='grib'\n                         parameter='tp').to_xarray()\ncml.load_dataset(\"s2s-ai-challenge-training-output-reference\",\n                         date=[20200102,20200109],\n                         parameter='tp').to_xarray()\n```\n\nSee the demo notebooks here: https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/notebooks.\n\nAccessing the training data :\n- Netcdf [nbviewer](https://nbviewer.jupyter.org/github/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/notebooks/demo_netcdf.ipynb) [colab](https://colab.research.google.com/github/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/notebooks/demo_netcdf.ipynb)\n- Grib [nbviewer](https://nbviewer.jupyter.org/github/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/notebooks/demo_grib.ipynb) [colab](https://colab.research.google.com/github/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/notebooks/demo_grib.ipynb)\n\n\nGetting the observation (reference output) dataset see the [demo_observations notebook](https://nbviewer.jupyter.org/github/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/notebooks/demo_observations.ipynb).\n\nGetting the benchmark dataset see the [demo_benchmark notebook](https://nbviewer.jupyter.org/github/ecmwf-lab/climetlab-s2s-ai-challenge/blob/main/notebooks/demo_benchmark.ipynb).",
    "bugtrack_url": null,
    "license": "Apache License Version 2.0",
    "summary": "Climetlab external dataset plugin for the S2S AI competition organised by ECMWF",
    "version": "0.9.0",
    "project_urls": {
        "Homepage": "https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge"
    },
    "split_keywords": [
        "meteorology"
    ],
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "blake2b_256": "c8eb288466bbf139d84222f920a5f8ef2db271089964679eb85dd4b727a1133f",
                "md5": "af5fd1301a090b6efe0d53fa39c17122",
                "sha256": "3ff0ef9d2c9df1282feb601e31f8c4d6cf21a6b142b056b7680038a2ebb17f28"
            },
            "downloads": -1,
            "filename": "climetlab-s2s-ai-challenge-0.9.0.tar.gz",
            "has_sig": false,
            "md5_digest": "af5fd1301a090b6efe0d53fa39c17122",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 33407,
            "upload_time": "2023-09-13T08:11:43",
            "upload_time_iso_8601": "2023-09-13T08:11:43.594217Z",
            "url": "https://files.pythonhosted.org/packages/c8/eb/288466bbf139d84222f920a5f8ef2db271089964679eb85dd4b727a1133f/climetlab-s2s-ai-challenge-0.9.0.tar.gz",
            "yanked": false,
            "yanked_reason": null
        }
    ],
    "upload_time": "2023-09-13 08:11:43",
    "github": true,
    "gitlab": false,
    "bitbucket": false,
    "codeberg": false,
    "github_user": "ecmwf-lab",
    "github_project": "climetlab-s2s-ai-challenge",
    "travis_ci": false,
    "coveralls": false,
    "github_actions": true,
    "requirements": [],
    "tox": true,
    "lcname": "climetlab-s2s-ai-challenge"
}
        
Elapsed time: 0.11038s